AI Skill Report Card

Documenting Product and Pitch

B+78·Aug 20, 2026·Source: Web
13 / 15

Given raw product context (screens, flows, stack, goals), produce two aligned outputs from one source of truth:

  1. Technical doc set — architecture, data models, API assumptions, permissions, QA scenarios
  2. Pitch narrative — problem/solution, market, business model, roadmap, traction plan

Don't write these independently — derive both from the same module breakdown so they never contradict each other.

Recommendation
Add a second, fuller worked example showing actual output snippets (e.g., a sample module table filled in, a sample pitch slide bullet) rather than just describing what would happen.
14 / 15
Progress:
- [ ] Step 1: Gather product context (purpose, users, screens, stack, decisions made)
- [ ] Step 2: Identify audiences and required outputs (dev-facing vs investor-facing)
- [ ] Step 3: Decompose product into core modules
- [ ] Step 4: For each module, define flows, data, screens, backend logic, permissions, edge cases
- [ ] Step 5: Draft technical documentation set
- [ ] Step 6: Draft pitch deck narrative from same module data
- [ ] Step 7: Cross-check consistency between both outputs
- [ ] Step 8: Refine for clarity, credibility, and audience fit

Step 1: Gather Context

Collect (ask only if genuinely missing, otherwise infer from what's given):

  • Product purpose and target users
  • Existing screens / current implementation state
  • Technical stack (e.g., Android/Kotlin, backend, DB)
  • Business assumptions already made (monetization hints, market, competitors)

Step 2: Identify Audiences

Two tracks, always:

  • Developers — need specs precise enough to build/QA against
  • Investors/partners — need a narrative that sells the vision without technical noise

Never mix technical jargon into pitch content or vague marketing language into specs.

Step 3: Decompose Into Modules

Standard module set for rental/marketplace/SaaS-like products (adapt names to the actual product):

  • Onboarding & authentication
  • User profiles & roles
  • Core listing/discovery entity (e.g., properties, jobs, products)
  • Application/booking/request workflow
  • Payments
  • Documents
  • Messaging/communication
  • Notifications
  • Admin tools
  • Analytics

Step 4: Per-Module Definition

For every module, capture in this exact shape:

FieldContent
User flowStep-by-step from entry point to completion
Data modelEntities, key fields, relationships
ScreensList with purpose of each
Backend logicCore rules, state transitions
PermissionsWho can do what (role-based)
Edge casesFailure states, empty states, race conditions
DependenciesOther modules or external services required

Step 5: Technical Documentation Set

Assemble into these sections, in order:

  1. Product overview (1 paragraph, technical framing)
  2. System architecture (diagram description + component list)
  3. Feature specifications (per module, from Step 4)
  4. User flows (consolidated, cross-module)
  5. Data models (full schema-level entity list)
  6. Backend/API requirements (endpoints implied by flows, auth model)
  7. QA scenarios (edge cases from Step 4, per module)
  8. Risks (technical debt, scaling concerns, security)

Step 6: Pitch Deck Narrative

Standard slide sequence — do not reorder:

  1. Problem
  2. Solution
  3. Product (screenshots/flow highlights, non-technical)
  4. Market opportunity (size, target segment)
  5. Business model / monetization
  6. Competitive positioning
  7. Traction / current state
  8. Roadmap
  9. Team (placeholder if not provided)
  10. Ask / next milestones

Same module data as Step 4, but stripped of implementation detail — express as user value, not backend logic.

Step 7: Cross-Check

Verify: roadmap in pitch deck matches technical doc's future-module notes; monetization logic matches payment module's actual data model; no feature claimed in pitch that isn't supported (even partially) in the tech spec.

Step 8: Refine

Cut redundancy, ensure consistent terminology across both documents (same module names, same entity names), tighten pitch language to be outcome-focused not feature-focused.

Recommendation
The module list table (Step 4) and section templates are strong frameworks—consider adding a concrete filled-in example of one module fully fleshed out to anchor the abstract structure.
14 / 20

Example 1: Input: "Homygo has a listings screen, tenant application flow, and basic payment stub. Built in Kotlin, no backend yet."

Output:

  • Module breakdown: Onboarding, Profiles (tenant/landlord roles), Listings, Applications, Payments (stubbed), Documents (not yet built — flagged as roadmap item)
  • Technical doc flags payments and documents as "assumed backend, not implemented" with proposed API shape
  • Pitch deck marks Payments/Documents as "Q2 roadmap" under Traction/Roadmap slides — consistent with tech doc's flagged status

Example 2: Input: "Need investor deck only, skip technical docs."

Output: Skip Steps 5, 7 (partial). Still run Steps 3–4 internally to ground the narrative in real product structure, but output only the 10-slide pitch sequence from Step 6. Note in output that module analysis was done internally to ensure claims are grounded, even though it's not shown.

Recommendation
Tighten Steps 5/6 descriptions slightly; some repetition between workflow steps and best practices/pitfalls sections could be consolidated.
  • Always derive pitch content from the same module analysis as tech docs — this is the core value, not writing them separately
  • Use role-based permission tables even in early-stage products; investors read this as maturity signal
  • Flag unbuilt/assumed features explicitly in tech docs ("assumed," "not yet implemented") rather than presenting speculation as fact
  • Keep pitch deck feature descriptions to outcomes ("tenants apply in one tap") never implementation ("Kotlin coroutine-based async form submission")
  • Roadmap should appear in both documents but adapted in vocabulary — same timeline, different framing
  • Writing tech docs and pitch deck as if for two unrelated products — causes contradictions investors or engineers will catch
  • Letting monetization slides promise features that have no corresponding data model or backend logic in the tech spec
  • Overloading pitch deck with schema/API details — kills narrative momentum
  • Skipping edge cases/permissions in "early-stage" specs — these are cheap to define early and expensive to retrofit
  • Treating module list as fixed — always adapt names/scope to the actual product instead of forcing a generic template
0
Grade B+AI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
13/15
Workflow
14/15
Examples
14/20
Completeness
17/20
Format
14/15
Conciseness
12/15